Kimi K3

Specifications

No specifications recorded yet.

Latest developments

Exclusive: Paying for frontier AI models buys 4-month head start at 5x the cost Ars Technica · 2 days ago · 46 Cognition Releases SWE-2: A Kimi K3 Post-Trained Coding Model That Matches Fable 5.1 on FrontierCode at 64% Lower Cost MarkTechPost · 4 days ago · 29

French Investors Value China’s Moonshot AI, Maker of Kimi K3, at $30 Billion Trending Topics · 6 days ago · 7

Cognition's SWE-2 Product Hunt · 6 days ago · 46

Reasoning Prefills on a Few Open Models, v1.1 Gist · 1 week ago · 7

The Open Source AI Stack Together AI · 1 week ago · 17 Cambridge University spinout launches AI model "competitive" with OpenAI and Anthropic Tech.eu · 2 weeks ago · 6 Harvey Introduces Harvey Tenet: A Kimi K3 Base Post-Trained with Fireworks for Long-Horizon Legal Agent Work MarkTechPost · 3 weeks ago · 12

September 2026

◆ French investors via a pooled SPV invested in Moonshot AI ahead of its planned Hong Kong listing, valuing the company at $30 billion pre-money Funding

◆ Cognition launched SWE-2, a post-trained coding model built on Kimi K3 and optimized with reinforcement learning for lower cost and improved coding performance Model release

◆ US NSA, CISA, and FBI issued an advisory accusing six Chinese AI firms of bulk distillation of capabilities from US frontier models Security issue

◆ Commentary and analysis circulated on the competitive impact of open-weight AI models versus closed-source models, including discussion of proposed U.S. restrictions tied to K3 Other

• 15 Sep Exclusive: Paying for frontier AI models buys 4-month head start at 5x the cost
• 12 Sep Cognition Releases SWE-2: A Kimi K3 Post-Trained Coding Model That Matches Fable 5.1 on FrontierCode at 64% Lower Cost
• 10 Sep French Investors Value China’s Moonshot AI, Maker of Kimi K3, at $30 Billion
• 10 Sep Cognition's SWE-2
• 10 Sep Reasoning Prefills on a Few Open Models, v1.1
• 9 Sep The Open Source AI Stack
• 1 Sep Cambridge University spinout launches AI model "competitive" with OpenAI and Anthropic

August 2026

◆ DeepSeek released DeepSeek-V4-Flash (0731) and updated API pricing with peak/off-peak rates for its V4 models Pricing change

◆ Alibaba releases Qwen3.8-Max, a 2.4 trillion parameter multimodal AI model with coding and autonomous agent capabilities Model release

• 23 Aug Harvey Introduces Harvey Tenet: A Kimi K3 Base Post-Trained with Fireworks for Long-Horizon Legal Agent Work
• 13 Aug DeepSeek V4 Pro: Rock-bottom Cost Per Task, But Trailing Kimi K3
• 6 Aug Baseten on Hugging Face Inference Providers 🔥
• 3 Aug Is memory the moat?
• 3 Aug LWiAI Podcast #253 - Opus 5, Gemini 3.6, Kimi K3, Hugging Face Hack
• 3 Aug Why Silicon Valley is divided over China’s powerful, cheap AI models
• 3 Aug Cogent AI Team Releases VR-1: A Frontier Cyber Reasoning Model That Composes and Verifies Enterprise Attack Paths
• 3 Aug China’s Alibaba takes another swipe at America’s AI supremacy
• 2 Aug Latest open artifacts (#23): Laguna S2.1, Inkling, & Kimi K3 show the utility of open models on the Pareto frontier
• 2 Aug The Sequence Radar #906: Last Week in AI: Open Models, Intelligent Robots, and the Price of Conviction
• 1 Aug Supabase Releases Evals: an Open Source Benchmark That Scores Claude Code, Codex and OpenCode on Real Supabase Tasks
• 1 Aug Kimi K3: The Complete Developer Guide

July 2026

◆ Moonshot released Kimi K3, a large open-weight AI model Open source release

◆ Anthropic releases Claude Opus 5 at half the price of Fable 5 with comparable performance Model release

◆ Moonshot AI releases Kimi K3, a 2.8 trillion parameter open-weight model achieving frontier-class performance Open source release

◆ Moonshot AI releases Kimi K3 open model with 2.8 trillion parameters and 1-million-token context window Open source release

◆ Moonshot AI releases Kimi K3, a 2.8 trillion parameter open-weight model with API pricing and delayed weights availability Model release

◆ Moonshot AI released Kimi K3, an open-weight large language model Model release

• 31 Jul Oxide and Friends: The Open Weight Revolution with Simon Willison
• 30 Jul Deploying Kimi K3 on Amazon SageMaker HyperPod and Amazon EKS
• 30 Jul Deploying Kimi K3 on AWS
• 30 Jul With Moonshot’s free Kimi K3, China changes the sovereign AI playbook
• 30 Jul Kimi K3 Redraws the Open Frontier, Muse Spark 1.1 Undercuts Competitors, Cloudflare Moves to Cut Off Crawlers
• 30 Jul Moonshot AI Open-Sources MoonEP: A Perfectly Balanced Expert Parallelism Library for MoE Training
• 29 Jul Claude Opus 5 became downright ruthless when tasked with running a vending machine
• 29 Jul Together AI announces strategic partnership with Moonshot AI to natively serve Kimi models
• 28 Jul “This is not in my top ten list of worries”: What Sam Altman thinks about model distillation
• 28 Jul Moonshot Releases Kimi K3 Model Weights on Hugging Face
• 28 Jul [AINews] Much ado about Open Weights
• 27 Jul moonshotai/Kimi-K3
• 27 Jul Kimi AI and kvcache-ai Open Sources ‘AgentENV’: A Distributed System that Powers Agentic Reinforcement Learning (RL) Training for Kimi K3
• 27 Jul Moonshot opens Kimi K3 weights — but few can run it
• 27 Jul Why China is giving away its best AI models
• 26 Jul Kimi K3 vs GPT-5.6 Sol on DeepSWE: Cost, Coding, and Routing
• 25 Jul Microsoft, Nvidia, Meta and 22 others defended open weights. Anthropic and OpenAI didn’t sign.
• 24 Jul Deep Learning Weekly: Issue 465
• 24 Jul ‘AI communism’, rogue models, and the why Kimi K3 spooked Wall Street
• 24 Jul Kimi K3 vs Claude Fable 5 on DeepSWE: Cost and Coding
• 23 Jul 🔮 Will Kimi K3 change the economics of AI?
• 23 Jul Experts say exploiting Anthropic’s Fable isn’t how Kimi K3 got so good
• 22 Jul Treasury threatens sanctions after White House claims Moonshot distilled Anthropic’s Fable
• 22 Jul Kimi K3: White House alleges Fable 5 siphoning
• 22 Jul Open models recap: more on Kimi K3, Qwen 3.8, Xi's WAIC speech, distillation, the open-closed gap, and what's next
• 21 Jul Moonshot launched Kimi K3. Then demand shut down subscriptions in 48 hours.
• 21 Jul US threatens sanctions against Chinese AI models over IP theft
• 21 Jul “Second only to Fable 5:” Alibaba talks the talk with Qwen3.8 without providing any real data
• 21 Jul Who's Afraid of Chinese Models?
• 21 Jul How to Build Complex Websites with Kimi K3 Prompts
• 21 Jul Washington Backed Away from Banning Chinese Open AI Models, For Now

Relationships

Products & technology

• Moonshot AI develops this model · 33 sources
• Moonshot develops this model · 17 sources
• AWS deploys this model · 2 sources
• Together AI deploys this model · 2 sources
• SWE-2 derived from this model · 2 sources
• Derived from Kimi K2 · 1 source
• Derived from Claude · 1 source
• Derived from Kimi K2.6 · 1 source
• Andon Labs deploys this model · 1 source

Competition

• Competes with Claude Fable 5 · 1 source
• Competes with OpenAI · 1 source

Investment & ownership

• Moonshot AI owns this model · 1 source